Macromolecules 1991,24, 5481-5483 5481

"Inverse" Organic-Inorganic Composite Materials. 2.' Free-Radical Routes into Nonsbrinking Sol-Gel Composites

We,1*2as well as others? have been interested in exploring the reverse side of traditional glass-fiber-reinforced, or- ganic polymer composites4 by developing methods for uniformly embedding organic polymers within inorganic glass matrices. Although "inverted", the basic composite principles remain intact: but now the advantages (and occasionally synergistic) relationship can result from the combination of a high-modulus organic polymer (providing high tensile strength) or an amorphous elastomer (pro- viding improved impact resistance) with a three-dimen- sionally cross-linked, inorganic matrix (providing high compressive strength). The realization of these inverted organic-inorganic materials requires formation of the inorganic matrix under conditions in which the organic polymers will survive. We herein report the synthesis of optically transparent, simultaneous interpenetrating com- positess through the synchronous formation of both the inorganic and organic components. This is accomplished by the hydrolysis and condensation of tetraalkenyl or- thosilicates (the sol-gel p r o ~ e s s ) ~ to form an inorganic Si02 matrix while simultaneously eliminating unsaturated al- cohols which are polymerized in situ using free-radical techniques. This simultaneous approach utilizing PO- lymerizable alkoxides circumvents the insolubility problem often associated with organic polymers in the sol-gel pre- curser solutions, while a t the same time it eliminates the ubiquitious shrinkage problem associated with drying of sol-gel derived xerogel materials.

The sol-gel process for the preparation of glasses under mild conditions has received much attention with respect to the formation of highly uniform monolithic glasses and inorganic ceramic composite^.^^^ The sol-gel process is based on the homogeneous hydrolysis and condensation of metal alkoxides in the presence of cosolvents to form highly cross-linked, solvent-swollen networks. Under controlled drying conditions, the excess water, liberated alcohol, and cosolvents can slowly be removed to yield large-scale, optically transparent monolithic samples. The simplest sol-gel process is the formation of ramified Si02 gels from the hydrolysis of tetraethyl orthosilicate (TEOS) (eq 1).

The mild conditions offered by the sol-gel process make it attractive for application in the formation of composite materials.lg The most common approach involves dis- solving a preformed organic oligomer or polymer in the sol-gel solution and then allowing the hydrolysis and condensation of the inorganic network to proceed. Under the appropriate conditions, the polymer remains uniformly embedded within the inorganic gel throughout the syn- thesis and drying steps.

In spite of its attractive features, the application of the sol-gel process in the synthesis of new composite materials is limited by the insolubility of many important engineering polymers within the sol-gel solution. In addition, the extraordinary shrinkage which occurs upon drying the

0024-9297/91/2224-5481$02.50/0

Table I Candidate Polymerizable Monomers for Nonshrinking

Composites

I1 I V

Scheme I

4 =AcrylatePolymer = LOW Density Si02

Network

solvent-swollengels (shrinkages of > W 7 5 9% are ~ o m m o n ) ~ precludes most molding processes and introduces con- siderable stress within these materials. The former problem can be circumvented by the in situ formation of both the inorganic and organic components.1*2*htf In order to address the latter problem, we have synthesized a series of tetraalkenyl orthosilicates possessing polymerizable groups (Table I) in place of the standard ethoxide or meth- oxide groups commonly used in the sol-gel p roce~s .~ The hydrolysis and condensation of these siloxane derivatives to form the inorganic network liberate a polymerizable alcohol. In the presence of the appropriate catalyst and by using a stoichiometric amount of water and the cor- responding alcohol as cosolvent, all components of these derivatives are polymerized. Since both the cosolvent and the liberated alcohol polymerize, gel drying is unnecessary and no gel shrinkage occurs. This overall process is illustrated in Scheme I.

Siloxane derivatives I-IV (Table I) are readily synthe- sized in high yields (typically >95 % ) by allowing the free alcohols to react with S ic4 in the presence of a base.1° Modifying the alkoxide groups in siloxane sol-gel pre- cursors is not without potential complications. Specifi- cally, the rate of hydrolysis is known to be inversely related to the steric bulk of the alkoxide moieties." We have therefore studied the NaF (10-25 pM) catalyzed hydrol- ysis kinetics of siloxane I1 with excess DzO in acetone-ds or DMSO-de and have measured a pseudo-first-order rate constant of 14.4 s-l for this derivative a t 27 OC. This value compares favorably with the hydrolysis rate we measured for the simple ethoxide derivative, TEOS, under similar conditions.

The organic free-radical polymerization of siloxane I1 can be initiated photochemically by using UV light (medium-pressure Hanovia Hg lamp through quartz) with or without azobis(isobutyronitri1e) (AIBN), thermally (>60 O C under Nz), ammonium persulfate (APS), and sodium metabisulfite as redox initiators a t room temperature and finally with APS and N,N,N',N'-tetramethylethylenedi-

0 1991 American Chemical Society

5482 Communications to the Editor

1

Macromolecules, Vol. 24, No. 19, 1991

lymerization of both organic and inorganic components displays synergistic, nonadditive behavior.

Materials with even greater moduli can be prepared by incorporating varying amounts of cross-linking agents into the organic phase. The composites formed as described above from the synchronous polymerization of tetravi- nylalkoxy orthosilicates are semiinterpenetrating networks composed of linear organic polymers and the three- dimensional Si02 network. These materials can be con- verted into simultaneous interpenetrating three-dimen- sional cross-linked networks (SIPNs) by incorporating small amounts of divinyl monomers into the reaction mixture prior to polymerization. For example, cross- linking of the organic component of materials formed from the reaction of I1 can be accomplished by using ethylene diacrylate (EDA) (5:l mole ratio II/EDA, 20:l vinyl/di- vinyl moieties). The dynamic storage modulus of these fully cross-linked SIPNs increases rapidly and enters the glassy regime within minutes of initiation.

3 P E

Y obl : x

1 m'~imu~taneo~s Condensation and

1 Free Radical

1 ""- Condensation

1 0 0 v --

0 I 2 0 0

Time (min)

Figure 1. Dynamic storage modulus, G' (at 5% strain), as a function of time for the independent free-radical polymerization of HEA (Free Radical), the independent hydrolysis condensation of siloxane I1 (Condensation), and the simultaneous condensation and free-radical polymerization of I1 (Simultaneous Condensation and Free Radical).

amine (TMEDA) at room temperature. The latter system is highly efficient and provides the greatest control and is therefore the preferred method. Quantitative polym- erization of the acrylate moieties of I1 is obtained under actual composite-forming conditions using the APS/ TMEDA initiator as evidenced by IR spectroscopy. In preliminary experiments, we discovered that TMEDA is a potent catalyst for sol-gel condensation; thus, it influ- ences the rate of polymerization of the organic and the inorganic components of the system. As a result, polym- erization of each monomer type is not independently initiated/catalyzed; this makes kinetic studies difficult. Nevertheless, it is possible to Yune" both polymerizations to proceed at similar rates by varying the amount of APS present in the system. Matching the two rates is important for the formation of optically transparent composites where scattering losses must be minimized. Systems with inorganic condensation rates much greater than the organic polymerization rates produce opaque, brittle glasses that shrink (due to the evaporation of unreacted monomer), while uncontrolled polymer precipitation (phase separa- tion) occurs in systems with fast (dominant) organic po- lymerization rates.

The actual inorganic condensation and hydrolysis process and the organic polymerizations under reaction conditions can be studied both independently and simul- taneously by measuring the dynamic storage modulus, G', as a function of time (Rheometrics RMS-705 mechanical spectrometer configured in a parallel-plate geometry). The modulus-time plots for the independent free-radical po- lymerization of 2-hydroxyethyl acrylate (HEA) (labeled Free Radical; ambient temperature, 1.2 equiv of HzO relative to HEA, 6 mM APS and 31 mM TMEDA in HEA), the independent hydrolysis/condensation of siloxane I1 (labeled Condensation; ambient temperature, 2.7 equiv of HzO relative to 11,21 mM NaF, 29 mM TMEDA in 11), and the simultaneous hydrolysis/condensation and free- radical polymerization of I1 (ambient temperature, 2.8 equiv of HzO relative to 11,18 mM NaF, 7.6 mM APS, 29 mM TMEDA in 11) are shown in Figure 1.

As can be seen in Figure 1, poly-HEA under these bulk polymerization conditions forms a gel which displays the behavior expected of a relatively weak hydrogen-bonded three-dimensional network. Likewise, the nascent inor- ganic matrix has a low modulus. In contrast, however, the composite material obtained from the simultaneous po-

Acknowledgment. We gratefully acknowledge finan- cial support for this work from UC Berkeley "Startup Funds"; the Office of Naval Research; The Center for Advanced Materials, Materials Science Division, Lawrence Berkeley Laboratory; E.I. du Pont Nemours and Co.; the Exxon Foundation; and Chevron Research. C.D. ac- knowledges the Eastman Kodak Co. for a Graduate Student Fellowship.

References and Notes

(1) Part 1: Ellsworth, M. E.; Novak, B. M. J. Am. Chem. SOC. 1991, 113, 2756.

(2) Novak, B. M.; Ellsworth, M.; Wallow, T.; Davies, C. Polym. Prepr. (Am. Chem. SOC., Diu. Polym. Chem.) 1990,31,698.

( 3 ) For selected references, see: (a) Wilkes, G. L.; Huang, H.; Carl- son, J. G. Polymer 1989,30,2001. (b) Wilkes, G. L.; Orler, B.; Huang, H. Polym. Prepr. (Am. Chem. SOC., Diu. Polym. Chem.) 1985, 26, 300. (c) Huang, H.; Orler, B.; Wilkes, G. L. Polym. Bull. 1985, 14, 557. (d) Huang, H.; Orler, B.; Wilkes, G. L. Macromolecules 1987,20,1322. (e) Pope, E. J. A.; Asami, M.; MacKenzie, J. D. J. Mater. Res. 1989,4,1018. (0 Schmidt, H. J. Non-Cryst. Solids 1989,112,419. (g) Philipp, G.; Schmidt, H. J. Non-Cryst. Solids 1984,63,283. (h) Glaser, R. H.; Wilkes, G. L. Polym. Bull. 1988, 19, 51. (i) Huang, H.; Glaser, R. H.; Wilkes, G. L. Polym. Prepr. (Am. Chem. SOC., Diu. Polym. Chem.) 1987,28,434. (j) Wilkes, G. L.; Huang, H.; Glaser, R. H. In Silicon Based Polymer Science; Zeigler, J. M., Fearon, F. W. G., Eds.; Advances in Chemistry Series 224; American Chemical Society: Washington, DC, 1990. (k) Mark, J. E.; Ning, Y. D.; Jiang, C. Y.; Tang, M. Y.; Roth, W. C. Polymer 1985,26, 2069.

(4) Lubin, G., Ed. Handbook of Composites; Van Nostrand Re- inhold Co.: New York, 1982.

(5) (a) Gerstle, F. P. In Encyclopedia of Polymer Science and Engineering; Kroschwitz, J. I., Ed.; John Wiley & Sons: New York, 1985; Vol. 3, p 776. (b) Manson, J. A.; Sperling, L. H. Polymers Blends and Composites; Plenum Press: New York, 1976.

(6) The interpenetrating network nomenclature used herein con- forms to that introduced for polymer-polymer systems. See: Sperling, L. H. Interpenetrating Polymer Networks and Related Materials; Plenum Press: New York, 1981. The actual phase behavior and morphology of these materials are unknown and still under investigation. Scanning electron microscopy (SEMI studies (JEOL-35C instrument a t 20 kV) on these composite materials show them to be highly uniform, as they exhibit no discernible phase separation at the ca. 1O2-nm level.

(7) (a) Ulrich, D. R. CHEMTECH 1988, 242. (b) Brinker, C. J.; Scherer, G. W. Sol-Gel Science; Academic Press: New York, 1990.

Macromolecules, Vol. 24, No. 19, 1991 Communications to the Editor 5483

(a) Mackenzie, J. D., Ulrich, D.,R. Eds. Ultastructure Processing of Advanced Ceramics; Wiley-Interscience: New York, 1988. (b) Klein, L. C., Ed. Sol-Gel Technology for Thin Films, Fibers, Preforms, Electronics, and Speciality Shapes; Noyea Publi- cations: Park Ridge, NJ, 1988. (c) Wright, A. F., Dupuy, J., Eds. Glass ... Current Issues; Martinus Nijhoff Publishers: Boston, 1985. Yoldas, B. E. J. Mater. Sci. 1986, 21, 1087. Ebelman, J. J. Justus Liebigs Ann. Chem. 1846,57, 331. (a) Aelion, R.; Loebel, A.; Eirich, F. J. Am. Chem. SOC. 1950, 72,5705. (b) McNeill, J. A.;DiCaprio, J. A,; Walsh,D. A.; Pratt, R. F. J. Am. Chem. SOC. 1980, 102, 1859.

(12) Alfred P. Sloan Fellow, 1991-93; Presidential Young Investi-

Bruce M. Novak'J) and Caroline Davies Department of Chemistry

University of California at Berkeley and the Center for Advanced Materials

Materials Sciences Division, Lawrence Berkeley Laboratory Berkeley, California 94720

Received April 25, 1991 Revised Manuscript Received July 16, 1991

gator, 1991-96.